Several authors have now addressed themselves to the dynamics of party identification. The meaning of such dynamics has, however, been left in some doubt. In particular, it is unclear whether changes in party attachments are best described as completely random phenomena or as phenomena exhibiting patterned time variation. This paper argues that the answer to the question depends in part on the observable consequences of changes in partisan ties. The analysis utilizes SRC panel data to describe patterns of variation in party identification and to attempt to assess their consequences for our understanding of the floating vote.

No abstract

No abstract

No abstract

No abstract